Low-Complexity Distributed Cell-Specific Bias Calculation for Load Balancing in UDNs

The challenge of the fifth generation (5G) of wireless systems is to overcome fundamental limits of existing cellular networks by means of advanced technologies such as network densification. Nevertheless, dense and random deployment of the network infrastructure poses new challenges in the network management. In particular, rethinking cell association policies and introducing new specific load balancing strategies will be mandatory to shift traffic from heavy loaded cells to light loaded neighbors. Toward this goal, cell biasing is considered one of the most effective way. This paper proposes a low complexity distributed cell-specific bias algorithm suitable for an actual implementation on highly dense scenarios. This solution takes into account the users' data rate requests and aims at minimizing the outage-rate defined as the amount of unsatisfied data rate requests. Numerical results show that despite its low complexity, the solution achieves performance only slightly lower than the benchmark near-optimum solution.